TypeError: 'NoneType' object cannot be interpreted as an integer

[Key-NE]

System Info

from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_chroma import Chroma
from langchain_ollama import OllamaEmbeddings

from langchain_ollama import ChatOllama

from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate

import os

from langchain.document_loaders import TextLoader,Docx2txtLoader

from langchain_community.document_loaders import TextLoader, Docx2txtLoader
from PyPDF2 import PdfReader

from langchain_core.runnables import RunnablePassthrough
from langchain_community.vectorstores import FAISS

from transformers import AutoModelForCausalLM, AutoTokenizer,BitsAndBytesConfig
import torch

os.environ["http_proxy"] = "http://127.0.0.1:11434"
os.environ["https_proxy"] = "http://127.0.0.1:11434"

#加载pdf文件并返回文本段
def load_single_pdf(file_path):
pdf_reader = PdfReader(file_path)
if not pdf_reader:
return None
ret = ''
for i,page in enumerate(pdf_reader.pages):
txt = page.extractText()
if txt:
ret += txt
return ret

#文档加载器
def data_loader(directory_path=r'E:\llm\deepseek\document_hc'):
all_docs = []
for root, dirs, files in os.walk(directory_path):
for file in files:
file_path = os.path.join(root, file)
if file.endswith('.txt'):
loader = TextLoader(file_path, encoding='utf-8')
elif file.endswith('.docx'):
loader = Docx2txtLoader(file_path)
elif file.endswith('.pdf'):
loader = load_single_pdf(file_path)
else:
continue
docs = loader.load()
all_docs.extend(docs)
return all_docs

将文本拆分为 docs文档

def split_text(txt,chunk_size=1000,overlap=100,directory_path =r'E:\llm\deepseek\document_hc'):
if not txt:
return None
splitter = RecursiveCharacterTextSplitter(chunk_size = chunk_size, chunk_overlap = overlap)
docs = splitter.split_documents(data_loader(directory_path))

# docs = splitter.split_text(txt)
return docs

def create_embeddings():
embeddings = OllamaEmbeddings(model="bge-m3:567m")
return embeddings

def load_llm_ollama(model_path_name = "deepseek-r1:7b"):
chat_model = OllamaEmbeddings(model=model_path_name)
return chat_model

使用Embeddings嵌入模型将文档保存到向量知识库存储

def create_vector_store(docs, embeddings, store_path):
# vectorstore = Chroma.from_documents(documents=split, embedding=local_embeddings)
# vector_store = FAISS.from_texts(docs,embeddings)

vector_store = FAISS.from_documents(docs, embeddings)
vector_store.save_local(store_path)
return vector_store

从文件加载向量知识库

def load_vector_store(store_path,embeddings):
if os.path.exists(store_path):
vector_store = FAISS.load_local(
store_path,
embeddings = embeddings,
allow_dangerous_deserialization=True
)
return vector_store
else:
return None

加载或者创建向量知识库

def load_or_create_vector_store(store_path,doc_file_path):
embeddings = create_embeddings()
vector_store = load_vector_store(store_path,embeddings)
if not vector_store:
docs = split_text(doc_file_path)
vector_store = create_vector_store(docs,embeddings,store_path)
return vector_store

从vector store查询上下文

def query_vector_store(vector_store, query, k=4, relevance_threshold=0.8):
similar_docs = vector_store.similarity_search_with_relevance_scores(query, k=k)
related_docs = list(filter(lambda x: x[1] > relevance_threshold, similar_docs))
context = [doc[0].page_content for doc in related_docs]
return context

Load llm(deepseek) and tokenizer

def load_llm(model_path,CUDA_Device):
quant_config = BitsAndBytesConfig(load_in_8bit=True)
model = AutoModelForCausalLM.from_pretrained(model_path,
device_map=CUDA_Device,
torch_dtype=torch.float16,
quantization_config=quant_config)
model = model.eval()

tokenizer = AutoTokenizer.from_pretrained(model_path,
    use_fast=False)
tokenizer.pad_token = tokenizer.eos_token

return model, tokenizer

def ask(model, tokenizer, prompt, CUDA_Device,max_tokens=512 ):
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids('<|eot_id|>')
]
input_ids = tokenizer([prompt],
return_tensors='pt',
add_special_tokens=False).input_ids.to(CUDA_Device)
generated_input = {
'input_ids': input_ids,
'max_new_tokens': max_tokens,
'do_sample': True,
'top_p': 0.95,
'temperature': 0.9,
'repetition_penalty': 1.1,
'eos_token_id': terminators,
'bos_token_id': tokenizer.bos_token_id,
'pad_token_id': tokenizer.pad_token_id
}

generated_ids = model.generate(**generated_input)
ans = tokenizer.decode(generated_ids[0], skip_special_token=True)
return ans

def main():
# 初始化
doc_file_path = r'E:\llm\deepseek\document_hc'
# store_path = './Data/Aquila.faiss'
store_path = r'E:\llm\deepseek\Data_vecstore\Aquila.faiss'
Embedding_Model = 'bge-m3:567m'
LLM_Model = r'E:\llm\deepseek\DeepSeek-R1-Distill-Qwen-1.5B'
# LLM_Model_path = r'E:\llm\deepseek\DeepSeek-R1-Distill-Qwen-7B'
CUDA_Device = 'cuda:0'

vector_store = load_or_create_vector_store(store_path, doc_file_path)
model, tokenizer = load_llm(LLM_Model,CUDA_Device)

while True:
    qiz = input('请输入您的问题: ')
    if qiz == 'bye' or qiz == 'exit':
        print('Bye~')
        break


    # Query context from vector store based on question, and compose prompt
    context = query_vector_store(vector_store, qiz, 6, 0.75)
    if len(context) == 0:
        # No satisfying context is found inside vector store
        print('无法从保存的向量存储中找到限定的上下文,在没有上下文的情况下与LLM交谈')
        prompt = f'请回答以下问题: \n{qiz}\n'
    else:
        context = '\n'.join(context)
        prompt = f'基于以下上下文: \n{context}\n请回答以下问题: \n{qiz}\n'

    ans = ask(model, tokenizer, prompt,CUDA_Device)[len(prompt):]
    print(ans)

if name == 'main':
main()

Who can help?

No response

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• My own task or dataset (give details below)

Reproduction

from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters import RecursiveCharacterTextSplitter

from langchain_chroma import Chroma
from langchain_ollama import OllamaEmbeddings

from langchain_ollama import ChatOllama

from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import ChatPromptTemplate

import os

from langchain.document_loaders import TextLoader,Docx2txtLoader

from langchain_community.document_loaders import TextLoader, Docx2txtLoader
from PyPDF2 import PdfReader

from langchain_core.runnables import RunnablePassthrough
from langchain_community.vectorstores import FAISS

from transformers import AutoModelForCausalLM, AutoTokenizer,BitsAndBytesConfig
import torch

os.environ["http_proxy"] = "http://127.0.0.1:11434"
os.environ["https_proxy"] = "http://127.0.0.1:11434"

#加载pdf文件并返回文本段
def load_single_pdf(file_path):
pdf_reader = PdfReader(file_path)
if not pdf_reader:
return None
ret = ''
for i,page in enumerate(pdf_reader.pages):
txt = page.extractText()
if txt:
ret += txt
return ret

#文档加载器
def data_loader(directory_path=r'E:\llm\deepseek\document_hc'):
all_docs = []
for root, dirs, files in os.walk(directory_path):
for file in files:
file_path = os.path.join(root, file)
if file.endswith('.txt'):
loader = TextLoader(file_path, encoding='utf-8')
elif file.endswith('.docx'):
loader = Docx2txtLoader(file_path)
elif file.endswith('.pdf'):
loader = load_single_pdf(file_path)
else:
continue
docs = loader.load()
all_docs.extend(docs)
return all_docs

将文本拆分为 docs文档

def split_text(txt,chunk_size=1000,overlap=100,directory_path =r'E:\llm\deepseek\document_hc'):
if not txt:
return None
splitter = RecursiveCharacterTextSplitter(chunk_size = chunk_size, chunk_overlap = overlap)
docs = splitter.split_documents(data_loader(directory_path))

# docs = splitter.split_text(txt)
return docs

def create_embeddings():
embeddings = OllamaEmbeddings(model="bge-m3:567m")
return embeddings

def load_llm_ollama(model_path_name = "deepseek-r1:7b"):
chat_model = OllamaEmbeddings(model=model_path_name)
return chat_model

使用Embeddings嵌入模型将文档保存到向量知识库存储

def create_vector_store(docs, embeddings, store_path):
# vectorstore = Chroma.from_documents(documents=split, embedding=local_embeddings)
# vector_store = FAISS.from_texts(docs,embeddings)

vector_store = FAISS.from_documents(docs, embeddings)
vector_store.save_local(store_path)
return vector_store

从文件加载向量知识库

def load_vector_store(store_path,embeddings):
if os.path.exists(store_path):
vector_store = FAISS.load_local(
store_path,
embeddings = embeddings,
allow_dangerous_deserialization=True
)
return vector_store
else:
return None

加载或者创建向量知识库

def load_or_create_vector_store(store_path,doc_file_path):
embeddings = create_embeddings()
vector_store = load_vector_store(store_path,embeddings)
if not vector_store:
docs = split_text(doc_file_path)
vector_store = create_vector_store(docs,embeddings,store_path)
return vector_store

从vector store查询上下文

def query_vector_store(vector_store, query, k=4, relevance_threshold=0.8):
similar_docs = vector_store.similarity_search_with_relevance_scores(query, k=k)
related_docs = list(filter(lambda x: x[1] > relevance_threshold, similar_docs))
context = [doc[0].page_content for doc in related_docs]
return context

Load llm(deepseek) and tokenizer

def load_llm(model_path,CUDA_Device):
quant_config = BitsAndBytesConfig(load_in_8bit=True)
model = AutoModelForCausalLM.from_pretrained(model_path,
device_map=CUDA_Device,
torch_dtype=torch.float16,
quantization_config=quant_config)
model = model.eval()

tokenizer = AutoTokenizer.from_pretrained(model_path,
    use_fast=False)
tokenizer.pad_token = tokenizer.eos_token

return model, tokenizer

def ask(model, tokenizer, prompt, CUDA_Device,max_tokens=512 ):
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids('<|eot_id|>')
]
input_ids = tokenizer([prompt],
return_tensors='pt',
add_special_tokens=False).input_ids.to(CUDA_Device)
generated_input = {
'input_ids': input_ids,
'max_new_tokens': max_tokens,
'do_sample': True,
'top_p': 0.95,
'temperature': 0.9,
'repetition_penalty': 1.1,
'eos_token_id': terminators,
'bos_token_id': tokenizer.bos_token_id,
'pad_token_id': tokenizer.pad_token_id
}

generated_ids = model.generate(**generated_input)
ans = tokenizer.decode(generated_ids[0], skip_special_token=True)
return ans

def main():
# 初始化
doc_file_path = r'E:\llm\deepseek\document_hc'
# store_path = './Data/Aquila.faiss'
store_path = r'E:\llm\deepseek\Data_vecstore\Aquila.faiss'
Embedding_Model = 'bge-m3:567m'
LLM_Model = r'E:\llm\deepseek\DeepSeek-R1-Distill-Qwen-1.5B'
# LLM_Model_path = r'E:\llm\deepseek\DeepSeek-R1-Distill-Qwen-7B'
CUDA_Device = 'cuda:0'

vector_store = load_or_create_vector_store(store_path, doc_file_path)
model, tokenizer = load_llm(LLM_Model,CUDA_Device)

while True:
    qiz = input('请输入您的问题: ')
    if qiz == 'bye' or qiz == 'exit':
        print('Bye~')
        break


    # Query context from vector store based on question, and compose prompt
    context = query_vector_store(vector_store, qiz, 6, 0.75)
    if len(context) == 0:
        # No satisfying context is found inside vector store
        print('无法从保存的向量存储中找到限定的上下文,在没有上下文的情况下与LLM交谈')
        prompt = f'请回答以下问题: \n{qiz}\n'
    else:
        context = '\n'.join(context)
        prompt = f'基于以下上下文: \n{context}\n请回答以下问题: \n{qiz}\n'

    ans = ask(model, tokenizer, prompt,CUDA_Device)[len(prompt):]
    print(ans)

if name == 'main':
main()

Expected behavior

Traceback (most recent call last):
File "E:\llm\deepseek\demo_test\langchain_0402.py", line 199, in
main()
File "E:\llm\deepseek\demo_test\langchain_0402.py", line 195, in main
ans = ask(model, tokenizer, prompt,CUDA_Device)[len(prompt):]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\llm\deepseek\demo_test\langchain_0402.py", line 159, in ask
generated_ids = model.generate(**generated_input)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Program Files\anaconda3\envs\LLM\Lib\site-packages\torch\utils_contextlib.py", line 116, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "D:\Program Files\anaconda3\envs\LLM\Lib\site-packages\transformers\generation\utils.py", line 2004, in generate
self._prepare_special_tokens(generation_config, kwargs_has_attention_mask, device=device)
File "D:\Program Files\anaconda3\envs\LLM\Lib\site-packages\transformers\generation\utils.py", line 1820, in _prepare_special_tokens
eos_token_tensor = _tensor_or_none(generation_config.eos_token_id, device=device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\Program Files\anaconda3\envs\LLM\Lib\site-packages\transformers\generation\utils.py", line 1817, in _tensor_or_none
return torch.tensor(token, device=device, dtype=torch.long)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: 'NoneType' object cannot be interpreted as an integer

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.